Fraud as a service

Abstract

A fraud detection as a service system is provided that can be implemented via one or more microservices that can be instantiated on an operation support system and/or business support systems. Each fraud detection microservice instantiated can have one or more fraud detection models and sets of rules for a particular type of incoming data stream or transaction and/or interaction data. In this way, there can be more than one fraud detection microservice operating on the OSS/BSS allowing each fraud detection microservice to be dynamically updated in real time, provide continuous integration and continuous delivery of services, and work with a particular data flow, providing for an optimal fraud detection process.

What is claimed is: 
     
       1. A system, comprising:
 interfacing with the first incoming transaction data prior to receiving a registration update from a registration component on the system; 
 instantiating the fraud detection microservice, at the registration component, as a virtual machine in a transactional processing system to monitor the first incoming transaction data for fraud; 
 in response to instantiating the fraud detection microservice, performing a clustering machine learning computation, a logistic regression machine learning computation, and a regression tree machine learning computation, resulting in clustered machine learning computations that have been performed by server equipment; 
 based on the clustered machine learning computations, determining that the fraud detection microservice has failed to detect a fraud event associated with a registration associated with the first incoming transaction; and 
 in response to the determining that the fraud detection microservice has failed to detect the fraud event and based on an inference generated based on artificial intelligence, updating the fraud detection microservice with a rule to mitigate the fraud event, wherein updating the fraud detection microservice comprises utilizing a machine-learning protocol to modify the rule to be applied to a second transaction represented by second transaction data, and wherein updating the fraud detection microservice further comprises classifying the second transaction as likely to be the fraud event. 
 
     
     
       2. The system of  claim 1 , wherein the operations further comprise:
 sending the fraud detection microservice with the rule based on the type of transaction. 
 
     
     
       3. The system of  claim 1 , wherein the operations further comprise:
 receiving second incoming transaction data associated with the second transaction, and wherein the second transaction is of a same type as the type of transaction. 
 
     
     
       4. The system of  claim 1 , wherein updating the fraud detection microservice is performed independently of maintaining the group of fraud detection microservices. 
     
     
       5. The system of  claim 1 , wherein the rule is a first rule, and wherein the operations further comprise:
 selecting a second rule from a group of rules associated with the type of transaction. 
 
     
     
       6. The system of  claim 1 , wherein the first incoming transaction is associated with a client identity of the transactional processing system. 
     
     
       7. The system of  claim 1 , wherein the type of transaction is a first type of transaction, wherein the operations further comprise:
 receiving third incoming transaction data associated with a third transaction, and wherein the third transaction is a second type of transaction. 
 
     
     
       8. The system of  claim 1 , wherein the operations further comprise:
 identifying the first incoming transaction data in the transactional processing system. 
 
     
     
       9. The system of  claim 1 , wherein a different fraud detection microservice of the group of fraud detection microservices is respectively instantiated for each type of incoming transaction data. 
     
     
       10. The system of  claim 1 , wherein the operations further comprise:
 implementing a new rule from an operator for subsequent incoming transaction data received after the first incoming transaction data. 
 
     
     
       11. The system of  claim 1 , wherein the first incoming transaction comprises an interaction between equipment associated with a client account and the transactional processing system. 
     
     
       12. A method, comprising:
 selecting, by a control panel interface of the fraud service equipment, the fraud detection microservice based on a type of the customer interaction data; 
 analyzing, by the fraud service equipment, the first data stream prior to receiving registration update data from a registration component; 
 enabling, by the fraud service equipment, the fraud detection microservice, at the registration component, using a virtual machine to monitor the first data stream for fraud; 
 in response to instantiating the fraud detection microservice, performing, by the fraud service equipment, a clustering machine learning computation, a logistic regression machine learning computation, and a regression tree machine learning computation, resulting in clustered machine learning computations; 
 based on the clustered machine learning computations, determining, by the fraud service equipment, that the fraud detection microservice has failed to detect a fraudulent activity; and 
 based on determining that the fraud detection microservice has failed to detect the fraudulent activity, utilizing, by the fraud service equipment, a machine-learning protocol to modify a rule to detect the fraudulent activity to be applied to a second data stream. 
 
     
     
       13. The method of  claim 12 , wherein the fraud detection microservice comprises a first rule based on a type of the customer interaction data. 
     
     
       14. The method of  claim 12 , wherein utilizing the machine-learning protocol comprises utilizing a classification process to determine that the second data stream is likely to be the fraudulent activity. 
     
     
       15. The method of  claim 12 , further comprising:
 receiving, by the fraud service equipment, a third data stream that comprises third customer interaction data that is associated with a different type of interaction data than the first customer interaction data. 
 
     
     
       16. The method of  claim 12 , wherein utilizing the machine-learning protocol does not change the group of fraud detection microservices. 
     
     
       17. The method of  claim 12 , wherein the fraud detection microservice is a first fraud detection microservice, and further comprising:
 implementing, by the fraud service equipment, a second fraud detection microservice to monitor incoming transaction data for the fraud. 
 
     
     
       18. A non-transitory machine-readable medium, comprising executable instructions that, when executed by a processor of a device, facilitate performance of operations, comprising:
 interfacing with first incoming transaction data prior to receiving a registration update from a registration component; 
 determining a first fraud detection microservice from a group of fraud detection microservices, accessible via server equipment, that is applicable to the first incoming transaction data; 
 in response to instantiating the first fraud detection microservice, performing a group of clustering machine learning computations, a group of logistic regression machine learning computations, and a group of regression tree machine learning computations, resulting in clustered machine learning computations; 
 based on determining that the first fraud detection microservice has failed to identify the fraudulent activity and based on an inference generated by artificial intelligence, utilizing machine-learning to update a second fraud detection microservice with a fraud detection rule to be applied to a second transaction type associated with second incoming transaction data, wherein utilizing the machine-learning comprises utilizing a classification process to determine that the second incoming transaction is threshold likely to be the fraudulent activity, and wherein utilizing the machine-learning is performed independently of performance the group of fraud detection microservices; 
 and instantiating, using the registration component, the first fraud detection microservice as a virtual machine and the second fraud detection microservice in the transactional processing system to monitor the first incoming transaction data and the second incoming transaction data for the fraudulent activity. 
 
     
     
       19. The non-transitory machine-readable medium of  claim 18 , wherein the operations further comprise:
 receiving third incoming transaction data associated with the first transaction type; and 
 adding an additional fraud detection rule to the first fraud detection microservice based on the third incoming transaction data. 
 
     
     
       20. The non-transitory machine-readable medium of  claim 18 , wherein the operations further comprise:
 selecting the second fraud detection microservice from the group of fraud detection microservices based on the second transaction type.